Sunday, February 9, 2014

British cryptologic security failures in WWII

Signals intelligence and codebreaking played an important role in WWII. British and American codebreakers solved many important Axis crypto systems, such as the German Enigma machine and the Japanese Navy’s code JN25.

Historians have not only acknowledged these Allied successes but they’ve probably exaggerated their importance in the actual campaigns of the war.
Unfortunately the work of the Axis codebreakers hasn’t received similar attention. As I’ve mentioned in my piece Acknowledging failures of crypto security all the participants suffered setbacks from weak/compromised codes and they all had some successes with enemy systems.

Britain, the Soviet Union and the United States did not have impenetrable codes. In the course of WWII all three suffered setbacks from their compromised communications.
Time to take a look at the British side and their worst failures.

Book cyphers
The basic British cryptosystem for important radio-traffic was the enciphered codebook. These 4-figure codebooks were enciphered with subtractor tables, using the non-carrying system. The military services had their own series of cyphers such as the War Office Cypher for the Army and the RAF cypher for the airforce plus there were diplomatic editions for the Foreign Office and the Interdepartmental Cypher that was used both by the services and the civilian organizations.

The codebook was basically a dictionary that assigned a 4-figure group to each word. For example the word ‘division’ would have the code 5538, ‘attack’ 2090, ‘artillery’ 0231 etc etc
So the cipher clerk would first use the codebook in order to find the code groups corresponding to the words of the message and then he would have to use the subtractor tables in order to encipher them. This means that each codegroup would be subtracted from the key groups (of the subtractor table) without carrying over the numbers.

For example let’s say that the following message is handed to the cipher clerk:
Enemy frigate sighted South-West of Malta.

Let’s assume that using the codebook this becomes: enemy=2591 , frigate=7482 , sighted=5556 , SW=3309 , Malta= 4610
So the message becomes 2591 7482 5556 3309 4610

This will be enciphered using the subtractor table valid for this time period. This was a book containing (usually) 15.000 numerical groups. The cipher clerk had to choose a random page and random starting point and then use that numerical sequence for enciphering the codegroups. Let’s say that the clerk chooses the 9th page and 2nd line from the subtractor table. He needs five key groups and these are 5668 8301 3496 3540 7778.
Now he would subtract the code groups of the message from these key groups, without carrying.

Enemy frigate sighted South-West of Malta   

2591 7482 5556 3309 4610 - code groups
5668 8301 3496 3540 7778 - key groups

3177 1929 8940 0241 3168 – cipher groups

The receiving party would identify from the indicator of the message the page and line that was used from the subtractor book and then subtract the received cipher sequence from the numerical key sequence. This would reveal the code groups whose meaning would be deduced from the Cypher book.
Obviously this was a time consuming operation and prone to errors due to mistakes in encipherment.

However it was thought at the time that the use of both a codebook and enciphering tables provided a very high level of security against enemy codebreakers.
Mistakes in the use of enciphered codebooks

The War Office Cypher and the Interdepartmental Cypher were read by the German codebreakers mainly thanks to physical compromise.
The Germans captured two copies of the WOC in 1940, one during the Norway campaign and the other near Dunkirk. By having the codebook and by taking advantage of frequent ‘depths’ (messages enciphered with the same numeric sequence) they were able to read traffic in the Middle East Theatre in the period summer 1941 till January 1942.

The official history ‘British intelligence in the Second World War’ vol2, p298 says: 
If under-estimation of the quality of Rommel's equipment was one reason why British confidence was high when the Crusader offensive began, another was the failure to allow for the efficiency of his field intelligence. By August 1941 the Germans were regularly reading the War Office high-grade hand cypher which carried a good deal of Eighth Army's W/T traffic down to division level, and they continued to do so until January 1942. Until then, when their success was progressively reduced by British improvements to the recyphering system, whereas GC and CS's success against the German Army Enigma continued to expand, this cypher provided them with at least as much intelligence about Eighth Army's strengths and order of battle as Eighth Army was obtaining about those of Rommel's forces.
The Interdepartmental Cypher was used by the Foreign Office, Colonial, Dominions and India offices and the Services. Also used by the Admiralty for Naval Attaches, Consular Officers and Reporting Officers. The basic book was captured from the British consulate in Bergen in May 1940, allowing the Germans to solve ‘depths’ during the period 1940-43. In their efforts they were assisted by poor British cipher practices. A security investigation in 1942 showed that the tables were overloaded, leading to heavy ‘depths’ and the indicators were not selected correctly.

The German exploitation of the ID Cypher ended on 15 June 1943 when the codebook was changed.
In both cases it is highly doubtful that the Germans could have solved this traffic on their own, without having the basic codebooks. It was a big mistake to continue using codebooks that were almost certainly expected to have fallen in enemy hands. Had the Brits introduced new versions in late 1940 or early 1941 the German success would have been nipped in the bud.

This was not however true in the case of the RAF Cypher. This was a 4-figure codebook enciphered with subtractor tables. The codebreakers of the Luftwaffe’s Chi Stelle did not have the basic book but they were able to ‘break’ into the system in March 1940 and traffic between the Air Ministry and Gibraltar, Malta, Habbaniya, Ismailia was read. From early 1941 to November 1942 traffic in the Med/Middle East was also compromised. According to Dr Voegele, chief cryptanalyst of the Luftwaffe: ‘From Sept. '41 to Nov. '42 the majority of the 200 - 400 daily intercepted 4 fig. messages could be decyphered with an average delay of 5 - 10 days, in single cases messages were decyphered the day of intercept.’
The German success was once again dependent on ‘depths’ and they didn’t have to search hard to find them. According to the report AIR 20/1531 ’R.A.F. signal communications: security’ in 1940-41 there were two series of enciphering tables in use with the RAF Cypher, the ‘Special’ for higher formations and the ‘General’ for all units. Traffic was not split evenly between the two tables because the units that had the ‘Special’ sets were usually equipped with the Typex cipher machine so they relied on that. Considering that the ‘General’ tables were the ones used all the time and that they were valid for three months it was only natural that there would be heavy ‘depths’, although the number of 150 given by the report is definitely impressive!

Thanks to the heavy ‘depths’ and the heavy traffic (up to 600 messages daily) the Luftwaffe cryptanalysts were able to solve new editions of the RAF Cypher in 1942. Their luck run out in November ’42 when the British security measures finally defeated them. Had the Brits introduced enciphering tables faster and thus avoided ‘depths’ it is difficult to see how the Germans could have succeeded in solving the RAF Cypher.

Naval codebooks
The Royal Navy’s main cryptologic systems were the Code and Cypher. The former was used for mid-level traffic and the latter for high level messages. Naval Cypher No1 (a 4-figure book) and the Administrative Code (a 5-figure book) were used from 1934 till August 1940. There was also the Auxiliary Code No3, (a 4-letter book) used from 1937 till August 1940 by small units. This was discontinued in August 1940 and instead small units used the Naval Code enciphered with Auxiliary Vessels tables.

It is immediately obvious that these systems were used for too long, thus compromising their security. If that wasn’t enough, instead of introducing new editions at the start of WWII they continued to be used up till August 1940!
These mistakes did not pass undetected by the codebreakers of Germany and Italy. The German naval codebreakers attacked these cryptosystems with success. In 1935-6 the Administrative Code was solved and in 1938 the Naval Cypher followed. At the start of the war the Auxiliary Code was also read with little difficulty. They were greatly aided in their efforts by the poor British decision to use the Administrative and Auxiliary codes unreciphered for non-confidential traffic. This allowed them to recover the true values of the codebooks and then focus only on breaking the enciphering tables.

Additionally there were mistakes in the construction of the codebooks and their use. According to the article ‘The Cryptographic Services of the Royal (British) and Italian Navies’, written by Admiral Luigi Donini, one of the top Italian codebreakers of WWII, some of the main British mistakes were:
1). Reuse of the same key sequence over and over, leading to ‘depths’.

2). The period of validity of the enciphering tables was badly commensurate with the level of radio traffic.
3). Numerical sequences from old table were sometimes added in the new ones. His explanation for this was ‘perhaps for quicker compilation’.

4). For many years the indicators were not enciphered thus making it very easy to locate ‘depths’. When enciphered indicators were first used the method was so clumsy that ‘it only caused us a two week crisis’.
5). The Cypher book did not contain homophones, ie did not assign two or more codegroups for a certain word of very high frequency.

6). Signals almost always began with the complete address to …from…
7). According to the article: ‘Geographical names were coupled, as an alternate interpretation, to vocabulary entries beginning with the same two or three letters. e.g.: 7184 = Give/Gibraltar, 0921 = Last/La Spezia, 4650 = Make/Malta, 2935 = All Concerned/Alexandria, 7714 = Left/Leghorn. This was of great help in identifying the geographical term when we had already identified the vocabulary term coupled thereto and vice versa;

Donini also points out that the failure to introduce new Code and Cypher books at the start of the war was the biggest mistake:
‘In my opinion, however, the most serious British carelessness or inadvertence was to keep in use for the whole first year of war with Germany the same principal naval cipher which was in force since the time of the Spanish civil war, broken by us (and by the Germans) in 1938.’

These failures meant that during most of WWII the Royal Navy’s most important cryptosystems were systematically exploited by the enemy with disastrous consequences for the Allies.
Disaster in the Atlantic – The case of Naval Cypher No3

Probably the biggest failure of British crypto security was the compromise of Naval Cypher No3, also known as the Convoy Cypher, since it was used in the Atlantic by the British, American and Canadian Navies.
In peacetime Britain was dependent on imports of raw materials and agricultural products from around the world. During WWII these shipments were vital for the smooth functioning of the war economy. The Germans knew that if they managed to sink the majority of supplies crossing the Atlantic then Britain would be economically strangled and would have no choice but to sue for peace. Even if this did not happen the lack of military supplies would make it impossible to launch heavy attacks on continental Europe.

The head of the U-boat service, Admiral Doenitz had calculated that with a large numbers of submarines he could achieve victory. His operational strategy was to overwhelm convoys with a large number of U-boats. These wolfpacks could evade the few escort vessels and sink the majority of merchant ships. The rest would disperse and could be picked off at a later time. In order for the wolfpack strategy to work the Germans needed to know the route and speed of the convoys in advance.
During the war the Germans had limited success in fielding a naval airforce and although they used traffic analysis and direction finding they had to rely mostly on codebreaking in order to track the Allied convoys.

According to TICOM report I-143 'Report on the Interrogation of Five Leading Germans at Nuremburg on 27th September 1945', p6
Doenitz stated emphatically that Sigint had been very valuable to him. It had been the best source of Naval Intelligence, and indeed, when air recce, etc., were not available, had often been the only source of operational information.

Since the convoys were vital for the Allied cause one would expect that every measure would be taken to ensure the impenetrability of their communications. Unfortunately the blunders of the responsible departments defy belief.
The need for a special codebook solely for Allied convoy duties had not been foreseen prior to WWII and thus the Brits had to give the Americans and Canadians copies of their Naval Cypher No3. The first problem was that their own codebook had to be used for longer than anticipated since Cypher No3 could not replace it as planned.

The security of the enciphered codebook system depended on the frequent changes of enciphering tables and the introduction of a new codebook after a reasonable period of time. Obviously the meaning of ‘reasonable’ changed during the war. The first Naval Cypher was used from 1934 till August 1940. The next edition Naval Cypher No2 was valid from August 1940 till January 1942. Its replacement Naval Cypher No4 (as No3 was used in the Atlantic) was valid from January 1942 till June 1943. So in the period 1940-1943 Cyphers were changed roughly every 1.5 years. One would expect that the Convoy Cypher would be valid for a similar timeframe or perhaps for security reasons it would have been changed sooner, maybe after 1 year.
Unfortunately Naval Cypher No3 wasn’t changed after 1 year. It wasn’t changed after 1.5 year. It was actually used from June 1941 till June 1943.

The continued use of the same basic codebook for two years meant that the codebreakers of the B-Dienst had an easy time recovering the true values and then they only needed ‘depths’ in order to read current traffic. The heavy traffic in the Atlantic combined with the small size of the enciphering tables (at 15.000 groups) led to heavy ‘depths’. For example M table-General: 218,000 groups in August ’42, S table-Atlantic: 148,000 groups in October 1942 and 220,000 in November.
The Germans would have to be incompetent in order to be defeated by a system that was used for two years and had huge ‘depths’ on a daily basis. They were able to ‘break’ into the traffic in December ’41, by February ’42 they had reconstructed parts of the book and till 15 December ’42 they were reading a large proportion of the traffic (at times up to 80% of intercepted messages). In December an indicator change set them back but from February ’43 they were again able to read the messages. Their greatest success with the Convoy Cypher was achieved in 1943. From February till June they often read signals 10-20 hours in advance of the actions mentioned in them. Also from February ‘42 to June ‘43 they could decode the daily Admiralty U-boat disposition signal nearly every day.

The German naval codebreakers were finally defeated in June ’43 when Naval Cypher No5 replaced No3 in the Atlantic. It didn’t take a miracle to end their success, just a new basic codebook that was long overdue….
Merchant Ships code

Since I’ve covered Naval Cypher No3 it’s important to also have a quick look at another very important system connected with the convoy battles. This was the cryptosystem used by merchant ships. Merchant ships had their own codebooks. During 1939 this was the International Code and Naval Appendix, from January ’40 to April ’42 it was the Merchant Navy Code and for the rest of the war the Merchant Ships Code. The Merchant Navy code and the Merchant ships code were captured from merchant ships and their enciphering tables were solved throughout the war. 
‘British intelligence in the Second World War’ vol2, p639 says that these two systems ‘were a prolific source of information to the B-Dienst second only to the Naval Cypher No3 in their importance to the battle of the Atlantic’.

Since these codebooks were in circulation around the world it’s not strange that they were physically compromised.  Once they had the basic book it was obviously easy to break into the traffic by taking advantage of ‘depths’. Could the British authorities have taken measures to secure this system? Yes and no. On the one hand a codebook used in such numbers, by ships around the world could not have been made secure. On the other hand a simple British countermeasure greatly limited the value of these decoded messages for the B-Dienst.
What was this measure? Simply enciphering the coordinates sent in the messages. This security measure was implemented in December 1943. Better late than never…

Cipher machines
The extensive use of enciphered codebooks by the British authorities does not imply that they were not aware of the benefits of a modern cipher machine. On the contrary had things gone according to plan they would have been supplied with one in large numbers and thus have been spared the losses that resulted from the fact that their communications were compromised.

In 1926, the British Government set up an Inter-Departmental Cypher Committee to investigate the possibility of replacing the book systems then used by the armed forces, the Foreign Office, the Colonial Office and the India Office with a cipher machine. It was understood that a cipher machine would be inherently more secure than the codebook system and much faster to use in encoding and decoding messages. Despite spending a considerable amount of money and evaluating various models by 1933 the committee had failed to find a suitable machine. Yet the need for such a device continued to exist and the Royal Air Force decided to independently fund such a project. The person in charge of their programme was Wing Commander Lywood, a member of their Signals Division. Lywood decided to focus on modifying an existing cipher machine and the one chosen was the commercially successful Enigma. Two more rotor positions were added in the scrambler unit and the machine was modified so that it could automatically print the enciphered text. This was done so these machines could be used in the DTN-Defence Teleprinter Network.
The new machine was called Typex (originally RAF Enigma with TypeX attachments). In terms of security it was similar to a commercial Enigma but had the additional security measure of multiple notches per rotor. This meant that during encipherment the rotors moved more often than in the standard Enigma machines.

The problems with Typex were:
1). Due to the failure of the Inter-Departmental Cypher Committee to select a cipher machine for mass production and the solitary efforts of the RAF in the mid 1930’s there were only a small number of Typex machines available at the start of WWII. The first contract in 1938 was for 350 machines and it’s doubtful that all would have been delivered by September 1939. Note that at that time the Germans had about 10.000 Enigmas in use.

2). The ability to print the enciphered text came at a heavy (literally!) price. While the German Enigma machine was relatively small and compact, the Typex version built in large numbers Typex Mk II was bulky and weighed 54kg. Thus it could only be used at prepared sites.

3). For the same reason Typex was too complex to mass produce during the war. According to ‘The Development of Typex’:

Some of the reasons for the low production rate are clear. Any rotor-based machine tends to be very complex mechanically. Figures 2 and 3 illustrate just how many different parts a Typex machine included. Typex Mk. VI contained about 700 parts, few of which were common to other models. Typex must have been a quartermaster's nightmare - much more so than Enigma, because of Typex's printer. Typex's relative complexity proved too much for the British machine tool industry. Overloaded as the industry was with the demands of the war economy generally, it took almost two years to obtain the machine tools required to manufacture Typex, despite the priority that would have been accorded to it. Only 2,300 Typex machines had been made by the end of 1942, 4,078 by December 1943 and 5,016 by May 1944.

For comparison’s purposes at least 40.000 Enigma machines were built by the Germans.
4). Because they were complex the machines often malfunctioned

For these reasons the Brits did not have, during WWII, a cipher machine in widespread use like the Germans did.
Low level codes

Does the compromise of low level codes count as a failure of cryptologic security? By their very nature low level codes are expected to secure information only for a limited amount of time. Still during WWII important information was passed on these systems and by reading these messages the Axis powers got order of battle data and even information on upcoming operations.
Some of the British low level codes extensively exploited by the Germans were the Syllabic cipher, Slidex, RAF Syko/Rekoh cards, Bomber code and the RN’s small ships codes.

The war diary of the German Army’s Inspectorate 7/VI has references to the Syllabic cipher, used at division level by the British Army. Although details on this system are lacking it is possible that it was similar to the well know Slidex.
The Slidex card system was used extensively by the British armed forces in the period 1943-45 but had very limited security. The decoded traffic provided the Germans with valuable intelligence, especially since it was used by ALO’s (Army Liaison Officers) in requesting air strikes.

Sykoh/Rekoh cards were used by the RAF as a low level system and they were extensively read by the codebreakers of the Luftwaffe. Messages from the planes of Coastal Command had important information.


The RAF also used the Bomber code, a daily changing 2-letter code table. During the Combined Bomber Offensive the RAF and the USAAF used the same tables each day with unfortunate consequences for at least one of them.

Small naval units used several systems (LOXO, FOXO, COFOX, MEDOX, TRAXO). These were extensively exploited by the B-Dienst.

The information was sometimes very important as in the run-up to the Normandy invasion.

Regarding low level codes it should be pointed out that the successes of General Rommel in N.Africa were definitely linked to the performance of his tactical signals intelligence unit NFAK 621, headed by Captain Seebohm. This unit relied on British low level codes, traffic analysis and direction finding and radio-telephone communications in order to monitor and anticipate British moves. Especially in conditions of mobile warfare, when messages were exchanged quickly with little regard for security, Seebohm’s men gave Rommel the edge.

Special Operations Executive codes
At the start of WWII the British foreign intelligence service SIS did not perform well. This led to the establishment of a similar organization in 1940, called SOE - Special Operations Executive. SOE was created to ‘set Europe ablaze’, so the main task was to organize resistance groups in the occupied countries and provide them with weapons and explosives. Still the fact that SOE had agents, spy groups and informers meant that it was trespassing on the activities of the SIS. As would be expected countless power struggles ensued between these two agencies.

The wartime performance of SOE was mixed at best. Although they certainly had their successes, countless SOE networks were compromised and their members arrested and executed. In Holland their entire network fell under German control in the famous Englandspiel operation.  In France they lost countless agents and networks. Just the fall of their Prosper network in 1943 led to the arrest of hundreds of resistance members.
SOE was disbanded in 1946 and most of its archives were destroyed postwar with some lost in a fire. Unfortunately the loss of the archives means that many questions about SOE wartime operations can never be answered.

Were some of the failures of SOE in Western Europe connected with their insecure cryptosystems? Leo Marks, head of the SOE cipher section, was constantly worried about the insecurity of their poem code but it took him till late 1943 to introduce the unbreakable letter one time pad. The change was gradual and even in 1944 many insecure systems continued to be used.
The main crypto system used by SOE for most of the war was double transposition, using a poem as a ‘key’ generator.

Why did SOE use this system for so long? In the beginning SOE communications were handled by SIS and it seems that the double transposition system was imposed by SIS. According to ‘Between Silk and Cyanide: The Story of SOE's Code War’, p40
Ozanne was advised by his friend and mentor Brigadier Gambier-Parry, C’s director of Signals, that their agents were going to continue using the poem-code (or some minor variation of it) as he had no doubt whatever that agents codes should be carried in their heads. This was all Ozanne needed to hear. What was good enough for the agents of the British Secret Service must be good enough for SOE.

and in page 57 when Marks wanted to change it he was told ‘Furthermore, I had greatly exaggerated the poem-code’s insecurity. Properly used it was perfectly suitable for SOE’s purposes.
Was the double transposition a secure and reliable system? Let’s see what a report titled ‘S.O.E. FIELD CIPHERS’ has to say:

Transposition systems based on poems or emergency phrases carried in the agent's head are a complete failure as the main system for S.O.E. type clandestine traffic.

1). Because if an agent is caught and tortured, he will almost certainly reveal the details of his poems – thus enabling the enemy to decipher all of his traffic which they have intercepted. This consideration is of paramount importance.

2). Low grade security is afforded when stereotyped messages are sent in transposition.
3). Under emotional stress the agent cannot remember his poems without difficulty, and when he can remember them he has not the time to use the code with the accuracy it requires.

4). After about 15-20 messages have been passed on a poem, the agent has a tendency to repeat the indicators he has previously used. If he is instructed to retain the list of indicators he will also retain his list of en clair messages; if he is instructed to destroy his messages, then it is a psychological certainty that he will revert to using indicators which have been tried and proven.

"If Indicator 123 has been cleared by the Home Station once, let us use it again and take no chances."
Leo Marks upgraded the security of SOE communications by introducing prepared keys for the double transposition system (called WOK’s-Worked-Out Keys) and with the Letter One time Pad- LOP. Even so it took a long time to replace the vulnerable systems with new ones and the report says:

‘S.O.E.'s field cipher security abroad was a mess from 1942 to the middle of 1943; from 1943 to 1945 it was put on a sound basis, but the rot was so embedded that some agents were using in 1945 conventions with which they had been issued in 1942.‘
Could SIS have helped out with SOE’s cryptologic problems?

According to ‘Between Silk and Cyanide: The Story of SOE's Code War’, p250 in early 1943 Leo Marks met Commander Dudley-Smith of Bletchley Park and demonstrated the Letter One time Pad- LOP system that he had devised. Dudley-Smith was impressed by the fact that Marks had thought of it by himself but also said: ‘As a matter of fact letter one time pads have been working very successfully for quite a long time’.
So SIS obviously knew how weak the double transposition system and probably did not use it themselves but they insisted that SOE should use it. For protecting their own communications they had LOP’s but they did not share this information with SOE. The strange behavior of the SIS leadership could be explained by their willingness to sabotage SOE operations, or at least to keep a close eye on them by forcing SOE to use a crypto system that could not resist a serious cryptanalytic attack.

If this interpretation is correct then the underhanded behavior of SIS definitely hurt the Allied war effort as some SOE communications were in fact decoded by Referat 12 (Referat Vauck) of the German Army’s signal intelligence agency Inspectorate 7/VI!
Securing British codes – 1943-45

Considering the information presented so far it is obvious that British cryptologic security was (more or less) a mess in the first half of WWII and continued to have serious vulnerabilities even in 1943-45.
However it should be acknowledged that the Brits systematically upgraded their systems and were finally able to secure their mid and high level codes in the period 1943-45. Did they succeed by investing huge resources on cipher security? Or by developing codes and ciphers so elaborate that no human mind could solve them? No, of course not. Simple security measures coupled with the new Stencil Subtractor system were enough to defeat the limited resources of the German codebreakers.

Book Cyphers were secured by introducing new versions and by enciphering them with the Stencil Subtractor Frame, a stencil that was used together with a daily changing numerical table. The SS Frame defeated ‘depths’ as the user could select different starting points for the enciphering sequence and that point was further enciphered.
The Royal Navy introduced new editions of the Code and Cypher every 6 months and it was expected that even this was not enough.

Without the basic codebooks and faced with the difficulty of locating depths, due to new stencil system, the German codebreakers were defeated.

In the case of cipher machines the Typex was upgraded with several sets of ‘split’ rotors, indicator books and a rewirable reflector. As Ralph Erskine puts it in ‘The Development of Typex’:
For the Germans to have been on an equal footing with Typex, as used by the British Army and the RAF, they would have had to find the wirings of from 120 to 252 rotors. Even Marian Rejewski or Alan Turing might have blanched at that Herculean task

Finally I’ve mentioned that SOE codes were upgraded with the Letter One time Pad, a system that if used correctly is unbreakable.
These improvements impressed the Germans and the report FMS P-038 'German Radio Intelligence’ says: ‘British radio communication was the most effective and secure of all those with which German communication intelligence had to contend’.

There was a downside however to all this. British cryptologic security was effective in this period but not efficient, in the sense that significant resources had to be invested in printing huge numbers of codebooks, enciphering tables, stencils, indicator books etc and transporting them to units across the globe. In the case of Typex the new procedures meant that cipher work fell to a fraction of the output achieved at the beginning of the war.


Allied successes in signals intelligence and codebreaking during WWII have received a lot of attention from Anglo-American historians. In Britain the successes of Bletchley Park are a source of national pride. The Germans had the Tiger tanks, their Me262 jet fighters and the V-2 rockets, the Americans were the ‘arsenal of democracy’ but the Brits had Bletchley Park.
However the many successes of Bletchley Park against Axis codes should not be used to distract from its failure to secure British codes. According to the article ‘Tunny Reveals B-Dienst Successes Against the ‘Convoy Code’:

GC&CS excelled at breaking the codes and ciphers of the Axis powers, and devoted huge resources to doing so. In March 1942, GC&CS employed about 1600 people on codebreaking operations, but only Travis (in theory) and Dudley-Smith were then assigned to investigating cipher security, even though Comsec was one of GC&CS’s two main functions. It was clearly too few, especially since Travis had no time to devote to Comsec, and Dudley-Smith was not a cryptanalyst. Even in October 1943, when GC&CS’s staff had more than trebled to over 4800, only Dudley-Smith (in a ‘part-time’ capacity!) and ‘two or three girls’ worked in the ‘Security of Allied Communications’ section, which investigated the security of the Army’s and Royal Air Force’s signals (and even those of some allies), in addition to the Royal Navy’s signals. Comsec is not as glamorous as codebreaking, but is probably more important.
Especially at sea the British mistakes prolonged the war. The Battle of the Atlantic was the longest continuous military campaign of World War II but it didn’t have to be. The only thing needed to stop the U-boat command was to secure the Convoy Cypher and here Bletchley Park undeniably failed.

It is said that ‘victory has a thousand fathers, but defeat is an orphan’. Perhaps that is the reason why the Allied failures of cryptologic security during WWII have not received the same attention as the ULTRA story. It is up to historians to correct this mistake.

Sources: various TICOM reports, ‘British Intelligence in the Second World War vol2’, ADM 1/27186Review of security of naval codes and cyphers 1939-1945’, AIR 20/1531 ’R.A.F. signal communications: security’, ‘European Axis Signal Intelligence in World War II’, FMS P-038 'German Radio Intelligence’, ‘Between Silk and Cyanide: The Story of SOE's Code War’, HS 7/41 ‘Section II: field cyphers; appendices L, M, N and O’ (from website, Intelligence and National Security article: Tunny Reveals B-Dienst Successes Against the ‘Convoy Code’, The Journal of Intelligence History article: ‘The Admiralty And Cipher Machines During The Second World War: Not So Stupid After All’, Cryptologia article: ‘The Cryptographic Services of the Royal (British) and Italian Navies’ , The Enigma Bulletin article: ‘The Development of Typex’, ‘Intelligence and Strategy: Selected Essays’, cryptomuseum
Acknowledgements: I have to thank Ralph Erskine for sharing information on British codes and ciphers, especially his article on Typex, the report AIR 20/1531 and the Naval Cypher No3’s S and M table statistics.

No comments:

Post a Comment